Two-phase flow measurement method based on differential pressure capacitance sensing and data knowledge driving

Abstract

The invention provides a two-phase flow measurement method based on differential pressure capacitance sensing and data knowledge driving. The two-phase flow measurement method comprises the steps that throat extension design is conducted on a standard Venturi tube; a multi-layer capacitive sensor array is embedded into a long-throat-neck Venturi tube, various transmitters are installed on the tube, and differential pressure, temperature and pressure data of the various transmitters are synchronously acquired by using a multi-channel data acquisition module; a configured capacitive sensor array excitation-measurement module is utilized to synchronously acquire capacitor array data; developing a gas-liquid two-phase flow calculation model based on data knowledge driving, and packaging the gas-liquid two-phase flow calculation model in a two-phase flow calculation module; capacitance, differential pressure, temperature and pressure data are input into a two-phase flow calculation module, and field learning is carried out by utilizing actually measured data, so that gas-liquid two-phase flow on-line real-time calculation is realized. The problem that time and space of data measured by traditional tandem type multiple sensors are inconsistent is solved, and the problem that when a traditional venturi tube is used for measuring two-phase flow, the phase volume fraction is lost is solved.

Description

Technical Field

[0001] The present invention relates to the technical field of multiphase flow measurement, and in particular to a two-phase flow measurement method based on differential pressure capacitance sensing and data knowledge driving. Background Art

[0002] Gas-liquid two-phase carbon dioxide (Carbon Dioxide, referred to as CO 2 ) is widely used in all aspects of carbon capture, utilization and storage (CCUS) projects. Its pipeline transportation is the key link between carbon capture, utilization and storage. Accurately measuring gas-liquid two-phase CO 2 The flow rate can provide an important measurement basis for my country's bulk carbon emission trading, and is of great significance to the safe and stable operation of CCUS projects and project cost estimation.

[0003] However, compared to other pipeline media (natural gas, oil and water), CO 2 The physical properties of CO are complex and changeable. Slight temperature or pressure fluctuations during transportation will cause CO 2 The phase state changes significantly, and most of them form gas-liquid two-phase CO 2 Mixed fluids, in addition, the introduction of impurity gases and load fluctuations caused by flexible power generation lead to pipeline CO 2 The flow rate changes dynamically in real time, making it difficult to accurately measure the flow rate using traditional methods. Domestic and foreign research shows that Coriolis flowmeters can directly measure CO 2 Mass flow, for single-phase gas, single-phase liquid or steady-state CO 2 Fluid measurement accuracy is high, but it is not suitable for gas-liquid two-phase or impurity CO 2 The flow measurement error is large. The existing commercial Coriolis meters can only be used in small pipe diameters and cannot meet the actual CCUS network pipeline (≥DN80). Ultrasonic and Venturi flowmeters are volume flowmeters. Only when the fluid density and content are known can CO2 be measured in large pipe diameters. 2 Accurate measurement of mass flow rate. The Venturi tube is less sensitive to density changes and is expected to measure CO2 with different pipe diameters and phases. 2 However, the measurement technology based on a single sensor principle has a limited measurement range and it is difficult to achieve accurate measurement of all parameters. Different sensor measurement data are seriously coupled with parameters such as flow rate and phase content. 2 Accurate flow measurement poses severe challenges. Summary of the invention

[0004] In order to overcome the shortcomings of the prior art, the purpose of the present invention is to provide a two-phase flow measurement method based on differential pressure capacitance sensing and data knowledge driven to achieve dynamic CCUS pipeline gas-liquid two-phase CO 2 Accurate measurement of flow rate.

[0005] To achieve the above object, the present invention provides the following solutions:

[0006] A two-phase flow measurement method based on differential pressure capacitance sensing and data knowledge driven, comprising:

[0007] The throat of the standard venturi tube is extended to obtain a long throat venturi tube.

[0008] embedding a multilayer capacitive sensor array into the long throat neck venturi;

[0009] Various transmitters are installed on the long throat neck venturi tube, and differential pressure, temperature and pressure data of various transmitters are synchronously acquired by using a multi-channel data acquisition module;

[0010] Configuring a capacitive sensor array excitation-measurement module for the multilayer capacitive sensor array to synchronously acquire capacitive array data;

[0011] Develop a data-knowledge driven gas-liquid two-phase flow calculation model, and encapsulate the gas-liquid two-phase flow calculation model in the gas-liquid two-phase flow calculation module;

[0012] The capacitance array data, the differential pressure, temperature and pressure data are input into the gas-liquid two-phase flow calculation module, and the measured data collected under field working conditions are used for on-site learning to achieve online real-time solution of the gas-liquid two-phase flow.

[0013] Preferably, the long throat neck venturi tube is built with a contraction section, an expansion throat section and an expansion section, the contraction section is located at the inlet of the long throat neck venturi tube, the expansion section is located at the outlet of the long throat neck venturi tube, and the expansion throat section is used to connect the contraction section and the expansion section; the contraction section and the expansion section are structures with gradual diameter changes; the expansion throat section is a straight tube structure; the length of the expansion throat section is greater than 10 times the diameter of the tube; pressure holes and temperature holes are respectively provided at both ends of the contraction section, the expansion throat section and the expansion section; the pressure holes and the temperature holes are connected to each of the transmitters; each of the transmitters is connected to the multi-channel data acquisition module.

[0014] Preferably, a multilayer capacitive sensor array is embedded in the long throat venturi tube, comprising:

[0015] Capacitive sensor arrays are installed on the contraction section, expansion throat section and expansion section of the long throat neck venturi tube respectively; the capacitive sensor array includes multiple layers of capacitive sensors; the capacitive sensors of the contraction section and the expansion section adopt a trapezoidal design, and the capacitive sensors of the expansion throat section adopt a rectangular design; for the capacitive sensor array, the number of capacitive sensors in each layer is 6, 8, 12 or 16, and multiple capacitive sensors are evenly distributed on the pipe wall of different sections, and a certain interval is maintained between adjacent capacitive sensors, and shielding rings of preset width are installed at both ends of each layer of capacitive sensors.

[0016] Preferably, the transmitter includes: a differential pressure transmitter, a temperature transmitter and a pressure transmitter; the differential pressure transmitter and the pressure transmitter share the same pressure tapping hole.

[0017] Preferably, the gas-liquid two-phase flow calculation model includes a flow pattern recognition network, a phase content calculation network and a gas-liquid two-phase total flow calculation network; the flow pattern recognition network is used to extract time domain features of preset multi-sensor measurement data to obtain time domain feature parameters, and convert the multi-sensor measurement data into a time domain feature matrix according to the time domain feature parameters, and use the convolutional neural network CNN-1 to construct a classification model between the time domain feature matrix and multiple flow patterns to achieve accurate identification of gas-liquid two-phase flow patterns; the time domain feature parameters are respectively mean value, standard deviation, variance, root mean square, peak value, root square amplitude, peak factor, waveform factor, pulse factor, margin factor, skewness and kurtosis; the multi-sensor measurement data include capacitance array data, differential pressure, temperature and pressure data; the phase content calculation network includes a convolutional neural network CNN-2 and a fully connected neural network DNN-1; wherein the output of the convolutional neural network CNN-2 is used as the fully connected neural network DNN-1 Partial input; the convolutional neural network CNN-2 is used to realize the effective mapping of the capacitance array data to the cross-sectional gas content, the input data of the convolutional neural network CNN-2 is the capacitance matrix eigenvalue sequence after the capacitance array data is normalized, and the capacitance matrix eigenvalue sequence is used as the partial input of the fully connected neural network DNN-1 by using the flattening layer; the fully connected neural network DNN-1 is used to map the input parameters related to the phase content in the multi-sensor measurement data to the mass gas content, the input data of the fully connected neural network DNN-1 are the gas-liquid density ratio, the dynamic viscosity ratio, the multi-differential pressure fluctuation coefficient, the multi-differential pressure ratio value and the output data of the convolutional neural network CNN-2, wherein the output data of the convolutional neural network CNN-2, the gas-liquid density ratio and the dynamic viscosity ratio are the input parameters for fitting the mass gas content using the capacitance array data, and the multi-differential pressure fluctuation coefficient and the multi-differential pressure ratio value are the input parameters for fitting the mass gas content using the multi-differential pressure data.

[0018] Preferably, the calculation formula of the gas-liquid two-phase total flow calculation network is: Among them, Wtp represents the total mass flow rate of gas-liquid two-phase flow; ε represents the thermal expansion coefficient; C represents the outflow coefficient; A T represents the cross-sectional area of ​​the extended throat section of the long throat neck venturi tube; ρ g and ρ l Represent the gas phase and liquid phase fluid density respectively; ΔP con represents the differential pressure generated by the gas-liquid two-phase fluid flowing through the contraction section of the long throat and neck venturi tube; x represents the mass gas content; β is the diameter ratio of the inlet section and the throat of the long throat and neck venturi tube; Φ g and Φ l Represent the gas phase and liquid phase false height correction coefficients respectively. The fully connected neural network DNN-2 is used to construct a nonlinear mapping model from the multi-differential pressure data of the long throat neck venturi to the gas phase or liquid phase false height. The input parameters of the model are multi-differential pressure ratio, gas-liquid density ratio, LM parameter and mass gas content. When the gas phase or liquid phase false height is obtained, combined with the obtained mass gas content, the total flow rate of the gas-liquid two-phase flow can be solved according to the above formula.

[0019] According to the specific embodiments provided by the present invention, the present invention discloses the following technical effects:

[0020] The present invention provides a two-phase flow measurement method based on differential pressure capacitance sensing and data knowledge driving, comprising: extending the throat of a standard Venturi tube to obtain a long throat neck Venturi tube; embedding a multilayer capacitance sensor array into the long throat neck Venturi tube; installing various transmitters on the long throat neck Venturi tube, and using a multi-channel data acquisition module to synchronously obtain differential pressure, temperature and pressure data of various transmitters; configuring a capacitance sensor array excitation-measurement module for the multilayer capacitance sensor array to synchronously obtain capacitance array data; developing a gas-liquid two-phase flow calculation model based on data-knowledge driving, and encapsulating the gas-liquid two-phase flow calculation model in a gas-liquid two-phase flow calculation module; inputting the capacitance array data, the differential pressure, temperature and pressure data into the gas-liquid two-phase flow calculation module, and using the measured data collected under field working conditions for field learning to achieve online real-time solution of the gas-liquid two-phase flow. The present invention overcomes the problem of time-space inconsistency of traditional serial multi-sensor measurement data, solves the problem of missing phase content in traditional venturi tube measurement of gas-liquid two-phase flow rate, realizes the deep integration and complementary advantages of two different sensing technologies, and provides a new method for gas-liquid two-phase CO2 in dynamic CCUS. 2 Key parameter calculation provides more effective multi-sensor fusion measurement data. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0022] Figure 1 A flow chart of a method provided by an embodiment of the present invention;

[0023] Figure 2 A schematic diagram of a measuring device provided in an embodiment of the present invention;

[0024] Figure 3 A flow chart of a measurement method provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0025] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0026] The purpose of the present invention is to provide a two-phase flow measurement method based on differential pressure capacitance sensing and data knowledge driving, which solves the problem of missing phase content when the traditional Venturi tube measures the flow of gas-liquid two-phase flow, realizes the deep integration and complementary advantages of two different sensing technologies, and provides a two-phase flow measurement method for gas-liquid two-phase CO in dynamic CCUS. 2 Key parameter calculation provides more effective multi-sensor fusion measurement data.

[0027] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0028] Figure 1 A flow chart of a method provided by an embodiment of the present invention, such as Figure 1 As shown, the present invention provides a two-phase flow measurement method based on differential pressure capacitance sensing and data knowledge driving, comprising:

[0029] Step 100: Extend the throat of the standard venturi tube to obtain a long throat venturi tube;

[0030] Step 200: embedding a multilayer capacitive sensor array into a long throat venturi tube;

[0031] Step 300: Install various transmitters on the long throat neck venturi tube, and use a multi-channel data acquisition module to synchronously obtain differential pressure, temperature and pressure data of various transmitters;

[0032] Step 400: configuring a capacitive sensor array excitation-measurement module for the multilayer capacitive sensor array to synchronously acquire capacitive array data;

[0033] Step 500: Developing a gas-liquid two-phase flow calculation model based on data-knowledge drive, and encapsulating the gas-liquid two-phase flow calculation model in a gas-liquid two-phase flow calculation module;

[0034] Step 600: Input the capacitance array data, differential pressure, temperature and pressure data into the gas-liquid two-phase flow calculation module, and use the measured data collected under field conditions for on-site learning to achieve online real-time solution of the gas-liquid two-phase flow.

[0035] like Figure 2 and Figure 3 As shown, the present invention provides a gas-liquid two-phase CO based on a long throat venturi tube and a capacitor plate array. 2 The flow measurement device and method are divided into the following steps:

[0036] Step 1: Design a long throat venturi sensor;

[0037] Step 2: Embed the Multilayer Capacitive Sensor Array into the Long Throat Neck Venturi

[0038] Step 3: Install various transmitters on the long throat venturi tube and configure a multi-channel data acquisition module to simultaneously obtain differential pressure, temperature and pressure data

[0039] Step 4: Configure the capacitive sensor array excitation-measurement module for the multilayer capacitive sensor array and synchronously acquire the capacitive array data

[0040] Step 5: Develop a data-knowledge-driven gas-liquid two-phase flow calculation model and encapsulate it in the gas-liquid two-phase flow calculation module

[0041] Step 6: Input the differential pressure, capacitance, temperature and pressure data into the gas-liquid two-phase flow calculation module, and through further on-site learning, realize the online real-time calculation of the gas-liquid two-phase flow.

[0042] Specifically, the detailed steps of this embodiment are as follows:

[0043] Step 1: Design a long throat and neck Venturi tube sensor. The gas-liquid two-phase flow process is complex and changeable, which makes it extremely difficult to accurately measure its flow rate. Conventional Venturi tubes can only obtain differential pressure data in the contraction section. The limited measurement data makes it difficult to accurately measure the gas-liquid two-phase flow rate. To address this shortcoming, the present invention extends the throat of the standard Venturi tube to synchronously measure the differential pressure of the contraction section, the extended throat section, and the expansion section to increase the amount of effective data and provide a rich data basis for the gas-liquid two-phase flow flow calculation. 2 Geometric parameters of the transport pipeline and gas-liquid two-phase CO 2 The flow characteristics of the pipe are suitable for gas-liquid two-phase CO 2 Long throat neck venturi tube sensor for fluid measurement. The long throat neck venturi tube is mainly composed of three parts: the contraction section, the expansion throat section and the expansion section. The contraction section and the expansion section belong to the gradual change structure of the pipe diameter, while the expansion throat section belongs to the straight pipe structure. The aperture ratio, contraction angle and expansion angle can be selected with reference to the design standards of conventional venturi tubes. In order to effectively extract the differential pressure information of the expansion throat, the length of the expansion throat is usually greater than 10 times the pipe diameter. Then, pressure and temperature holes are set at both ends of the contraction section, the expansion throat section and the expansion section, respectively, to obtain the differential pressure data of 3 different sections and the temperature and pressure data of 4 different positions, which are used for the subsequent solution of gas-liquid two-phase CO 2 Flow and density information.

[0044] Step 2: Embed the multilayer capacitive sensor array into a long throat venturi tube. When the long throat venturi tube is applied to the gas-liquid two-phase flow rate, it is important to consider the false height correction. However, the false height correction requires the phase holdup to be known in advance. 2 For non-conductive fluids represented by carbon monoxide, the capacitance method can be used to obtain the key intermediate parameter information of phase content. 2 The flow process is complex and changeable, and its flow rate and phase content change dynamically. There are also large differences in the phase content at different positions at the same time. In order to solve the problem of time and space inconsistency of traditional serial multi-sensor measurement data, the present invention directly embeds the capacitive array sensor into the surface of the long throat neck venturi tube, that is, the capacitive sensor array is installed in the contraction section, expansion throat section and expansion section of the long throat neck venturi tube, respectively, to achieve the deep integration and complementary advantages of two different sensing technologies, and to provide a reference for the gas-liquid two-phase CO2 sensing in dynamic CCUS. 2The key parameter solution provides more effective multi-sensor fusion measurement data. Since different sections of the long throat venturi tube have different pipe structures, the capacitive sensors of the contraction section and the expansion section adopt a trapezoidal design, while the capacitive sensors of the extended throat section adopt a rectangular design, and then wrap them outside the long throat venturi tube. For the capacitive sensor array, the number of capacitive sensors in each layer can be 6 / 8 / 12 / 16, and multiple capacitive sensors are evenly distributed on the pipe wall of different sections. A certain interval is maintained between adjacent capacitive sensors. Shielding rings of a certain width are installed at both ends of each layer of the capacitive sensor array. A certain pipe wall thickness is maintained between the capacitive sensor array and the measured fluid. The pipe wall material is usually made of polytetrafluoroethylene with a low relative dielectric constant. A shielding layer is added to the outer ring of the capacitor plate array sensor to ensure the quality of the measurement signal. The pressure resistance of the sensor is improved by glue filling to achieve gas-liquid two-phase CO in high-pressure pipelines. 2 Safe and stable measurement of fluids.

[0045] Step 3: Install various transmitters on the long throat venturi tube and configure a multi-channel data acquisition module to synchronously obtain differential pressure, temperature and pressure data. Install differential pressure, temperature and pressure transmitters on the pressure and temperature holes of the long throat venturi tube respectively. The differential pressure and pressure transmitters share the same pressure hole to ensure the consistency of the measured data. Different types of transmitters are mainly used to convert the physical quantities sensed by the long throat venturi tube into standard voltage or current output, which is convenient for the synchronous acquisition of multi-sensor measurement data. Usually, according to the gas-liquid two-phase CO 2 Parameters such as flow, temperature and pressure range are selected for key parameters such as the measurement range, working conditions, and output of differential pressure, temperature and pressure transmitters. Due to the large number of measurement parameters, in order to facilitate the synchronous acquisition of multi-sensor data, transmitters of the same output type are usually selected, such as two-wire 4-20mA standard current output. Then the multi-channel data acquisition module is connected to the differential pressure, temperature and pressure transmitters, and the electrical output signals are converted into digital signals. The differential pressure, temperature and pressure measurement data at different positions of the long throat neck venturi tube are synchronously obtained and used as the original data for gas-liquid two-phase flow calculation. In practical applications, according to the gas-liquid two-phase CO 2 The response time of the flow rate and differential pressure transmitters, and setting of appropriate sampling rates to ensure the dynamic CCUS medium pipe transmission of gas-liquid two-phase CO 2 Effective extraction of flow information.

[0046] Step 4: Configure the capacitive sensor array excitation-measurement module for the multilayer capacitive sensor array, and synchronously obtain the capacitive array data. For the multilayer capacitive sensor array, use signal lines to connect the individual capacitive sensors to the capacitive sensor array excitation-measurement module. Each capacitive sensor serves as both an excitation sensor and a measurement sensor, and the two types of functions are time-division multiplexed. For the single-layer capacitive sensor array, a sequential excitation-measurement working mode is used to obtain a full cycle of capacitive measurement data. For the capacitive sensor arrays of different layers, the same working mode is used, and the excitation and measurement are independent of each other to synchronously obtain the capacitance array measurement data of different segments. In practical applications, according to the gas-liquid two-phase CO 2 The flow rate range is set appropriately, and the excitation frequency and sampling frame rate are set to ensure that the gas-liquid two-phase CO in the dynamic CCUS pipeline is 2 Effective extraction of flow information.

[0047] Step 5: Develop a gas-liquid two-phase flow calculation model based on data-knowledge drive, and encapsulate it in a gas-liquid two-phase flow calculation module. The core of the gas-liquid two-phase flow calculation module lies in the flow calculation model. The present invention develops a gas-liquid two-phase flow calculation model based on data-knowledge dual drive according to the measurement characteristics of the differential pressure-capacitance integrated sensor, and then encapsulates it in a gas-liquid two-phase flow calculation module, which can work offline in real time. The calculation model proposed in the present invention mainly includes three parts: flow pattern identification network, phase content calculation network and gas-liquid two-phase total flow calculation network. First, a gas-liquid two-phase CO2 flow calculation model based on multiple differential pressure and capacitance sequences is constructed. 2 The flow pattern recognition network extracts time domain features from multi-sensor measurement data. The time domain feature parameters used are mean, standard deviation, variance, root mean square, peak, root square amplitude, peak factor, waveform factor, pulse factor, margin factor, skewness and kurtosis. Based on the preliminary feature extraction, the original measurement data can be converted into a time domain feature matrix. The convolutional neural network CNN-1 is used to build a classification model between the time domain feature matrix and multiple flow patterns to achieve gas-liquid two-phase CO 2Accurately identify the flow type. The following phase content and total flow calculation models are constructed under different flow types to overcome the problem that the measurement of gas-liquid two-phase flow parameters is greatly affected by the flow type. Secondly, based on formulas (4) and (8), a phase content calculation model based on the eigenvalue of the capacitance measurement matrix and the differential pressure fluctuation is constructed. This model consists of two parts: the convolutional neural network CNN-2 and the fully connected neural network DNN-1. The output of the convolutional neural network CNN-2 is used as part of the input of the fully connected neural network DNN-1. The convolutional neural network CNN-2 is mainly used to realize the effective mapping of capacitance measurement data to cross-sectional gas content. Its input data is the normalized capacitance matrix eigenvalue sequence, which is used as part of the input of the fully connected neural network DNN-1 using the flattening layer to avoid the problem that the convolutional neural network CNN-2 training cannot be completed due to the lack of cross-sectional gas content reference value. The fully connected neural network DNN-1 mainly maps the input parameters related to phase content in the two sensor measurement data to mass gas content. Its input data are gas-liquid density ratio, dynamic viscosity ratio, multi-differential pressure fluctuation coefficient, multi-differential pressure ratio value and output data of convolutional neural network. Among them, the output data of convolutional neural network CNN-2, gas-liquid density ratio and dynamic viscosity ratio can be regarded as input parameters for fitting mass gas content using capacitance measurement data, and multi-differential pressure fluctuation coefficient and multi-differential pressure ratio value can be regarded as input parameters for fitting mass gas content using multi-differential pressure data. Combining convolutional neural network CNN-2 with fully connected neural network DNN-1 realizes the effective fusion of multi-differential pressure and capacitance plate array data of long throat and neck venturi, solves the conversion problem between cross-sectional gas content and mass gas content, and realizes the accurate mapping of multi-sensor measurement data to mass gas content.

[0048] Finally, the gas-liquid two-phase CO is obtained using formula (1): 2 The total mass flow rate, the virtual height correction is the key to achieve accurate measurement of the total flow rate of gas-liquid two-phase flow. For gas-liquid two-phase flow dominated by liquid phase, the liquid phase virtual height is usually selected as the intermediate quantity for calculating the total mass flow rate; for gas-liquid two-phase flow dominated by gas phase, the gas phase virtual height is usually selected as the intermediate quantity for calculating the total mass flow rate. The fully connected neural network DNN-2 is used to construct a nonlinear mapping model from long throat and neck venturi multi-differential pressure data to gas phase or liquid phase virtual height. The input parameters of the model are multi-differential pressure ratio, gas-liquid density ratio, LM parameter and mass gas content. When the gas phase or liquid phase virtual height is obtained, combined with the mass gas content obtained in the previous step, the total flow rate of the gas-liquid two-phase flow can be solved according to formula (1).

[0049] In summary, the above three parts are all based on the measurement principle of differential pressure-capacitance integrated sensing to construct neural networks, integrating knowledge-driven information. Then, they are fully trained using measured data from different working conditions to realize the offline construction of the gas-liquid two-phase flow calculation model, and finally encapsulated in the gas-liquid two-phase flow calculation module.

[0050] Step 6: Input the differential pressure, capacitance, temperature and pressure data into the gas-liquid two-phase flow calculation module, and realize the online real-time solution of the gas-liquid two-phase flow through further on-site learning. Connect the multi-channel data acquisition module and the capacitive sensor array excitation-measurement module to the gas-liquid two-phase flow calculation module, and input the differential pressure, capacitance, temperature and pressure data into the pre-trained gas-liquid two-phase flow calculation model as the actual input data. At the same time, considering the working conditions of the actual application site, use certain measured data for on-site learning, optimize the flow calculation model, improve the robustness of the actual application of the flow calculation model, and improve the accuracy of the gas-liquid two-phase flow solution.

[0051] Furthermore, this embodiment derives the differential pressure, capacitance, temperature, pressure and other multi-sensor measurement data and the gas-liquid two-phase CO respectively according to the measurement principles of different sensing means and the fluid mechanics model. 2 Mechanistic model of flow parameters (see the following section for the above formula).

[0052] According to the Bernoulli equation, the differential pressure of the contraction section and the gas-liquid two-phase CO 2 The total flow relationship is,

[0053]

[0054] Where W tp It represents the total mass flow rate of gas-liquid two-phase flow, in kg / s; ε represents the thermal expansion coefficient; C represents the outflow coefficient; A T Indicates the expanded throat cross-sectional area of ​​the long throat venturi tube, in m 2 ρ g and ρ l Respectively represent the gas phase and liquid phase fluid density, unit kg / m 3 ; ΔP con It represents the differential pressure generated when the gas-liquid two-phase fluid flows through the contraction section of the long throat venturi tube, in Pa; x represents the mass gas content; β is the diameter ratio of the inlet section of the venturi tube to the throat; Φ g and Φ l Represent the gas phase and liquid phase false height correction coefficients respectively.

[0055] Domestic and foreign scholars have established a variety of classic false height correction formulas through theoretical analysis and experimental data. However, these false height correction formulas are mainly established under wet conditions, and none of them is applicable to gas-liquid two-phase CO 2Through comparative analysis, these false height correction models are closely related to operating parameters such as LM parameters and gas-liquid density ratio. Previous studies have shown that the multi-differential pressure ratio of the long-throat neck venturi tube is also strongly correlated with the false height and mass gas content. Therefore, the gas phase or liquid phase false height can be expressed as a multivariate nonlinear function of the long-throat neck venturi multi-differential pressure ratio, LM parameters and gas-liquid density ratio.

[0056]

[0057] In the formula, f OR The multivariate nonlinear function model representing the virtual height of the gas or liquid phase, ΔP con / ΔP str and ΔP str / ΔP exp They represent the differential pressure ratio between the contraction section and the expansion throat, and the differential pressure ratio between the expansion throat and the expansion section, respectively.

[0058] Compared with single-phase fluid, the differential pressure fluctuation characteristics of gas-liquid two-phase fluid are more complex, statistically have strong regularity, contain rich flow characteristics, and are of great significance for the measurement of gas-liquid two-phase flow parameters. For the multi-section differential pressure data of the long throat neck venturi tube, the differential pressure fluctuation coefficient can be used to quantitatively describe its dynamic fluctuation characteristics.

[0059]

[0060] Where N represents the data length of the differential pressure signal, ξ(·) represents the relative fluctuation of the dynamic differential pressure, that is, the degree of differential pressure fluctuation. represents the average differential pressure over a period of time, and i represents the number of the differential pressure signal sequence. The differential pressure fluctuations in different sections of the long throat neck venturi can be described by formula (3). Previous studies have shown that the differential pressure fluctuation coefficient is positively correlated with the mass gas content. As the mass gas content increases, the interaction between the gas and liquid phases increases, resulting in more intense differential pressure fluctuations, and the differential pressure fluctuation coefficient will also increase accordingly. Therefore, a mathematical model of multi-stage differential pressure fluctuation coefficient and mass gas content can be derived.

[0061]

[0062] In the formula, f MQ (·) represents the multivariate nonlinear function model of mass gas fraction.

[0063] The measurement data of the capacitor plate array is defined as a discrete mapping from the boundary excitation voltage to the boundary measurement voltage. Taking a single-layer 16-electrode capacitor plate array as an example, its capacitance matrix is ​​expressed as:

[0064]

[0065] In order to reduce the impact on the sensor structure parameters and the background dielectric constant in the measured field, and to enhance the difference of the capacitance matrix, it is normalized.

[0066]

[0067] In the formula, represents the normalized capacitance matrix, represents the reference capacitance matrix, that is, the capacitance matrix under uniform dielectric constant distribution. According to eigenvalue decomposition, the normalized capacitance matrix satisfies,

[0068]

[0069] In the formula, ω i and λ i denote the i-th eigenvector and its corresponding eigenvalue, respectively, and λ denotes the eigenvalue vector of the normalized capacitance matrix. Thus, the relationship between the eigenvalue sequence of the normalized capacitance matrix and the gas content of the cross section can be constructed.

[0070] α=f VF (λ) (8)

[0071] Where α represents the cross-sectional air content; f VF (·) represents the multivariate function model of mass gas fraction.

[0072] In summary, both the long-throat venturi tube and the capacitor plate array can measure phase holdup. The difference is that the mass gas holdup can be obtained by using the long-throat venturi multi-differential pressure ratio and fluctuation coefficient, and the cross-sectional gas holdup can be obtained by using the normalized capacitance matrix eigenvalue sequence. The two gas holds are not equal in most cases. In order to facilitate the effective fusion of multi-sensor measurement data, the slip ratio model proposed by Butterworth is used to convert the two phase holds.

[0073]

[0074] In the formula, μ g and μ l are the dynamic viscosities of the gas phase and the liquid phase, respectively, and a 1 、a 2 、a 3 and a 4 is an undetermined constant, which is usually determined by fitting with measured data. 2 The density of a fluid can be calculated based on the corresponding temperature and pressure.

[0075] The beneficial effects of the present invention are as follows:

[0076] (1) The present invention proposes a design scheme based on a differential pressure-capacitance integrated sensor, which breaks through the problem of temporal and spatial inconsistency of traditional serial multi-sensor measurement data, solves the problem of missing phase holdup in traditional venturi tube measurement of gas-liquid two-phase flow rate, and realizes the deep integration and complementary advantages of two different sensing technologies, providing a new method for gas-liquid two-phase CO2 sensing in dynamic CCUS. 2 Key parameter calculation provides more effective multi-sensor fusion measurement data.

[0077] (2) The present invention proposes a gas-liquid two-phase flow parameter solution method based on a data-knowledge driven model, which integrates differential pressure-capacitance integrated sensing to measure gas-liquid two-phase CO 2 The fluid mechanism model is introduced into the deep learning model as a physical prior constraint. The multi-physical field numerical simulation and experimental measurement results are used as a comprehensive data set to participate in the training and testing of the deep learning model. At the same time, the gas-liquid two-phase flow parameter solution model is further trained and optimized by field learning, which improves the performance of the deep learning model for gas-liquid two-phase CO 2 The interpretability and generalization of the artificial intelligence model for flow calculation solves the problem of poor adaptability of the key parameter measurement model of gas-liquid two-phase flow in practical applications, and realizes the gas-liquid two-phase CO 2 Accurate flow measurement.

[0078] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the various embodiments can be referenced to each other.

[0079] The principles and implementation methods of the present invention are described in this article using specific examples. The description of the above embodiments is only used to help understand the method and core idea of ​​the present invention. At the same time, for those skilled in the art, according to the idea of ​​the present invention, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting the present invention.

Claims

1. A two-phase flow measurement method based on differential pressure capacitance sensing and data knowledge driven, characterized in that: include: The throat of the standard venturi tube is extended to obtain a long throat venturi tube. embedding a multilayer capacitive sensor array into the long throat neck venturi; Various transmitters are installed on the long throat neck venturi tube, and differential pressure, temperature and pressure data of various transmitters are synchronously acquired by using a multi-channel data acquisition module; Configuring a capacitive sensor array excitation-measurement module for the multilayer capacitive sensor array to synchronously acquire capacitive array data; Develop a data-knowledge driven gas-liquid two-phase flow calculation model, and encapsulate the gas-liquid two-phase flow calculation model in the gas-liquid two-phase flow calculation module; The capacitance array data, the differential pressure, temperature and pressure data are input into the gas-liquid two-phase flow calculation module, and the measured data collected under field working conditions are used for on-site learning to achieve online real-time solution of the gas-liquid two-phase flow.

2. The two-phase flow measurement method based on differential pressure capacitance sensing and data knowledge driven according to claim 1 is characterized in that: The long throat neck venturi tube is equipped with a contraction section, an expansion throat section and an expansion section, the contraction section is located at the inlet of the long throat neck venturi tube, the expansion section is located at the outlet of the long throat neck venturi tube, and the expansion throat section is used to connect the contraction section and the expansion section; the contraction section and the expansion section are structures with gradual changes in pipe diameter; the expansion throat section is a straight pipe structure; the length of the expansion throat section is greater than 10 times the pipe diameter; pressure holes and temperature holes are respectively provided at both ends of the contraction section, the expansion throat section and the expansion section; the pressure holes and the temperature holes are connected to each of the transmitters; and each of the transmitters is connected to the multi-channel data acquisition module.

3. The two-phase flow measurement method based on differential pressure capacitance sensing and data knowledge driven according to claim 2 is characterized in that: A multilayer capacitive sensor array is embedded in the long throat venturi tube, comprising: Capacitive sensor arrays are installed on the contraction section, expansion throat section and expansion section of the long throat neck venturi tube respectively; the capacitive sensor array includes multiple layers of capacitive sensors; the capacitive sensors of the contraction section and the expansion section adopt a trapezoidal design, and the capacitive sensors of the expansion throat section adopt a rectangular design; for the capacitive sensor array, the number of capacitive sensors in each layer is 6, 8, 12 or 16, and multiple capacitive sensors are evenly distributed on the pipe wall of different sections, and a certain interval is maintained between adjacent capacitive sensors, and shielding rings of preset width are installed at both ends of each layer of capacitive sensors.

4. The two-phase flow measurement method based on differential pressure capacitance sensing and data knowledge driven according to claim 2 is characterized in that: The transmitter includes: a differential pressure transmitter, a temperature transmitter and a pressure transmitter; the differential pressure transmitter and the pressure transmitter share the same pressure taking hole.

5. The two-phase flow measurement method based on differential pressure capacitance sensing and data knowledge driven according to claim 2 is characterized in that: The gas-liquid two-phase flow calculation model includes a flow pattern recognition network, a phase content calculation network and a gas-liquid two-phase total flow calculation network; the flow pattern recognition network is used to extract time domain features of preset multi-sensor measurement data to obtain time domain feature parameters, and convert the multi-sensor measurement data into a time domain feature matrix according to the time domain feature parameters, and use the convolutional neural network CNN-1 to build a classification model between the time domain feature matrix and multiple flow patterns to achieve accurate identification of gas-liquid two-phase flow patterns; the time domain feature parameters are mean value, standard deviation, variance, root mean square, peak value, root square amplitude, peak factor, waveform factor, pulse factor, margin factor, skewness and kurtosis; the multi-sensor measurement data includes capacitance array data, differential pressure, temperature and pressure data; the phase content calculation network includes a convolutional neural network CNN-2 and a fully connected neural network DNN-1; wherein the output of the convolutional neural network CNN-2 is used as part of the fully connected neural network DNN-1 Input; the convolutional neural network CNN-2 is used to realize the effective mapping of the capacitance array data to the cross-sectional gas content, the input data of the convolutional neural network CNN-2 is the capacitance matrix eigenvalue sequence after the capacitance array data is normalized, and the capacitance matrix eigenvalue sequence is used as part of the input of the fully connected neural network DNN-1 by using the flattening layer; the fully connected neural network DNN-1 is used to map the input parameters related to the phase content in the multi-sensor measurement data to the mass gas content, the input data of the fully connected neural network DNN-1 are the gas-liquid density ratio, the dynamic viscosity ratio, the multi-differential pressure fluctuation coefficient, the multi-differential pressure ratio value and the output data of the convolutional neural network CNN-2, wherein the output data of the convolutional neural network CNN-2, the gas-liquid density ratio and the dynamic viscosity ratio are the input parameters for fitting the mass gas content using the capacitance array data, and the multi-differential pressure fluctuation coefficient and the multi-differential pressure ratio value are the input parameters for fitting the mass gas content using the multi-differential pressure data.

6. The two-phase flow measurement method based on differential pressure capacitance sensing and data knowledge driven according to claim 5 is characterized in that: The calculation formula of the gas-liquid two-phase total flow calculation network is: Among them, W tp represents the total mass flow rate of gas-liquid two-phase flow; ε represents the thermal expansion coefficient; C represents the outflow coefficient; A T represents the cross-sectional area of ​​the extended throat section of the long throat neck venturi; ρ g and ρ l Represent the gas phase and liquid phase fluid density respectively; ΔP con represents the differential pressure generated by the gas-liquid two-phase fluid flowing through the contraction section of the long throat and neck venturi tube; x represents the mass gas content; β is the diameter ratio of the inlet section and the throat of the long throat and neck venturi tube; Φ g and Φ l Represent the gas phase and liquid phase false height correction coefficients respectively.

Images